tsunami

log in
history

Al and Luis meeting notes

Luke Breuer
2012-03-20 00:55 UTC

overarching themes
  • decrease latency to below the "critical threshold", so that focus is not lost
things to talk about
notes
better generalized search syntax
For all of my ideas, I'll want something at least as powerful as Google's search operators. It'd be nice to have some sort of "intellisense", maybe if you type : at the beginning of a line or after a space.
refining
When I am learning a new topic, I feel like I have to do a lot of refining of raw material—the kind of refining that many have done in the past. And yet, current refined products are often undesirable: they might be too 'aggregated', biased, and/or lacking crucial information. My intuition says that there is an intermediate product which is still quite subjective, but could greatly reduce the time it takes to 'grok' a new topic.

The problem seems to be filtration. Either it is done ahead of time by people with their own agendas, or I have to do it myself—and risk missing important facts and theories if I cannot spend enough time on a thing. One way I avoid such holes is to cross-reference my knowledge base with friends. But this is an inherently limited process—there is a significantly nonzero cost for each friend I add to the "cross-referencing" list. Must this be how things work? Surely there is a way to digitally store who I trust and exactly how I trust them, and use this to find other people I may wish to consider trusting?

tl;dr Why can I not have like-minded friends—perhaps only like-minded on certain topics—share their knowledge and findings with me more transparently and efficiently?
precise linking
I want to be able to discuss anything. And I want to see when something is being discussed. This means websites and books. When I'm reading Immanuel Kant, it'd be nice to know what friends would be up for talking about him, but I'd also like to be able to connect with people who had interesting things to say about some specific passage!

Why can't I navigate between academic papers with the utmost of ease?
organic categorization and classification
Oftentimes, I don't know what form my data will take until I have some critical amount of it. Why spend the time and energy on finding a good way to structure data when you might only have a few instances of it?

discussion



* lots of information is just text
* hard to search
* hard to aggregate

currently available:
  1. sandy crude oil
  2. pre-molded plastic
--> nothing like gasoline, which is nice and intermediate and very versatile

forums idea
  • start with a forum where people are actively trying to accomplish something, instead of just wanting to understand some topic for fun
    • e.g., remote-control hobby forum (or hobby)
      • an advantage of this is that the users are fairly heterogenous; some will be experts in motors, another in video cameras, etc.
  • have the option to enforce certain thread structures, e.g. the philosophy group Luis sometimes attends (Socrates Cafe, with focus on collaboration over argumentation)
    question: topic of discussion
    proposed answer (with explanation)
    agree, or
    re-state proposed answer to ensure understanding was obtained, and/or ask question to clarify question,
    voice objections (prefer "elucidations", since the idea is that this is a collaboration instead of argument)

random stuff
  • heat maps
  • rotating text to, e.g. see the history
  • ontogeny recapitulates phylogeny — cutting the onion of complexity
    • how can I quickly see how an idea/concept has evolved through time
  • track what the internet browser viewport is spending time on
    • Al's "gonging" idea, where one can see at one points in a document readers tend to leave because they're no longer interested in the content

when Luis is trying to learn
  • books need to be more interactive
  • be able to talk to the book, or talk to the author, or another expert

"Better Information": Organically grown, tended by trusted individuals, attested to by experts, linked to _____.
  • I'm trying to come up with a nice one-sentence description that doesn't sound horrible.
  • with the "linked to ______", I'm trying to cover the "linked precisely" idea below

As structured as you want: generally, the more there is of it the more structure you'll want.
  • text-only might be acceptable in the beginning, but richer data types should be available on demand
linked precisely: if there is discussion of a specific term (or even numerical factor) in an equation, I should be able to tell the computer enough so that it can highlight that term and even let me click on the term in the equation and jump to the relevant discussion
  • a bad spot is PDFs, but I think there's software that lets you highlight arbitrary sections of PDFs. Whether or not one can, e.g. draw boxes around pieces of graphs (better: select one or more ranges & domains & data sets) is unknown.
discussed easily: using coarse to precise linking, so that I can, e.g. question whether the weight in a motor spec is correct
  • I want to see exactly what is being discussed by whom in discussions, on specifications of hardware/software,
trust others' choices about what is interesting/irrelevant/spam
  • instead of having a global "karma" or "reputation" for users, model this after the real world, where I trust various people's opinions, each more on some topics than other topics
    • I might trust the people my friends trust, but less — one could test
online identity will become important
  • a single person may have multiple identities


Questions to investigate:
* how did Wikipedia get started?




What happens when the page is so long that not everybody is looking at the same thing?





vvv old stuff vvv
Overarching Themes
  • decrease latency to below the "critical threshold", so that focus is not lost
Things to talk about

Notes
Trust Network
When I am learning a new topic, I feel like I have to do a lot of refining—the kind of refining that many have done in the past. Is there a way to do some refining beforehand and store that product, digitally? When I'm reading material, I feel like I'm consuming crude oil when I could be consuming spacecraft fuel.

The problem seems to be filtration. Either it is done ahead of time by people with their own agendas, or I have to do it myself—and risk missing important facts and theories if I cannot spend enough time on a thing. One way I avoid such holes is to cross-reference my knowledge base with friends. But this is an inherently limited process—there is a significantly nonzero cost for each friend I add to the "cross-referencing" list. Must this be how things work?

Why can I not have like-minded friends—perhaps only like-minded on certain topics—share their knowledge and findings with me more transparently and efficiently?
Precise Linking
I want to be able to discuss anything. And I want to see when something is being discussed. This means websites and books. When I'm reading Immanuel Kant, it'd be nice to know what friends would be up for talking about him, but I'd also like to be able to connect with people who had interesting things to say about some specific passage!

Why can't I navigate between academic papers with the utmost of ease?